From Tonga Tabu to Torres Strait. 103 



officers on board H.M.S. "Challenger" were surprised to 

 find that, beyond a depth varying between 1200 and 1400 

 fathoms, the thermometers ceased to register any further 

 decrease of temperature. The latter was observed to fall 

 from 25° C. at the surface to i°.8 C. between 1200 and. 1400 

 fathoms, and to remain stationary or nearly stationary from that 

 level down to the bottom. At Station 184, the bottom-tem- 

 perature in 1400 fathoms was i°.8 C. ; at Station 183, in 1700 

 fathoms, i°.7 C. ; at Station 182, in 2275 fathoms, i°.4 C. ; at 

 Station 176, between the Hebrides and Fiji, in 1450 fathoms, 

 2° C. ; and at Station 175, in 1350 fathoms, i°.8 C. The same 

 phenomenon was found to occur under similar conditions in the 

 seas of the Indian Archipelago, and tended to confirm the 

 opinion formed at the time — that its cause must be sought in 

 the partial or complete suspension of all deep-sea communica- 

 tion by intervening submarine ridges or submerged areas of 

 elevation. 



This circumstance affords one of the most convincing 

 proofs of the existence of a system, of thermal circulation in 

 a horizontal direction, which embraces the whole of the ocean 

 as well as the minor seas in proportion to the facilities of sttb- 

 marine communication. The almost uniform level of the 

 isotherms observed in landlocked basins also shows that the 

 obliquity or gradient of the isotherms stands in direct relation 

 to the presence or absence of currents of different temperature, 

 and therefore of different origin, and m^oving in different direc- 

 tions in the various areas which compose the aqueous envelope 

 of our planet. In the same manner, it has been explained on a 

 previous occasion that the gradients of the temperature-curve 

 stand in direct relation to the presefice or absence of currents of 

 different teinperature, origijz, and direction, in the various strata 

 between the surface and the bottom. 



An examination of Table IX. shows that, although the 



